The Kaczmarz algorithm proposed in [1] is one of the most effective and most computationally simple one-step estimation algorithms. In a number of subsequent studies the possibility of acceleration the Kaczmarz algori...
详细信息
adaptive algorithms have better optimization performance than stochastic gradient descent(SGD) in many *** is a widely used adaptive algorithm,but recent studies have shown that it suffer from non-convergence and poor...
详细信息
adaptive algorithms have better optimization performance than stochastic gradient descent(SGD) in many *** is a widely used adaptive algorithm,but recent studies have shown that it suffer from non-convergence and poor generalization *** [1] is also an adaptive algorithm,which is a variant of *** slightly modifying Adam ' s second moment,AdaX has achieved the generalization ability comparable to *** work aims to improve the AdaX algorithm with faster convergence speed and higher training *** has a momentum component,while Nesterov's accelerated gradient(NAG) is theoretically and experimentally superior to classical ***,we replace the classical momentum term of AdaX with NAG,and obtain the resulting algorithm named Nesterov's accelerated AdaX(Nadax).We conduct simulation experiments on MNIST,Fashion-MNIST,and Cifar-10 *** results show that Nadax significantly improves AdaX and is still competitive when compared with Adam's variants.
This paper describes an algorithm that is supposed to reduce the differences between the theoretical planned routes of commercial vehicles and the practical execution of routes based on historical validated data. The ...
详细信息
ISBN:
(纸本)9781665460156
This paper describes an algorithm that is supposed to reduce the differences between the theoretical planned routes of commercial vehicles and the practical execution of routes based on historical validated data. The resulted routes will not be as optimized as the they can be, but they should reflect more accurately the reality, keeping in mind local specific conditions, driver habits or other business factors that can not be used in a route planning algorithm. The first prototype of the algorithm use input parameters received from the historical execution of real-life vehicle routes. The reason for developing such an adaptive algorithm is to get smaller differences between planned and executed routes, a very important business indicator for commercial vehicle fleets. This paper is based on the work on Eureka project PN-III-P3-3.S-EUK-2019-0210.
Large-size workpieces in traditional industrial production have complex lighting environments that cannot meet the requirements for online detection. To improve the quality of images acquired under non-uniform illumin...
详细信息
Large-size workpieces in traditional industrial production have complex lighting environments that cannot meet the requirements for online detection. To improve the quality of images acquired under non-uniform illumination conditions. In this paper, an adaptive algorithm for non-uniform illumination in large-size workpieces online detection is proposed. First, a color space conversion is performed to preserve the color information, and then an improved effective guided filter is used to estimate the illumination information in the value channel, the illumination information is divided into the illuminance layer and the detail layer, 2D adaptive gamma function, and histogram specification are used for the illuminance layer to solve the illumination problem, and an optimized enhancement factor is used for the detail layer to enhance it. Finally, the enhanced image is obtained after fusion. By comparing with MSR, LIME, and FGif algorithms, the proposed algorithm can effectively solve the problem of non-uniform illumination in online detection, not only improving the image quality but also the operating speed that meets the requirements of online detection.
Approximation and uncertainty quantification methods based on Lagrange interpolation are typically abandoned in cases where the probability distributions of one or more system parameters are not normal, uniform, or cl...
详细信息
Approximation and uncertainty quantification methods based on Lagrange interpolation are typically abandoned in cases where the probability distributions of one or more system parameters are not normal, uniform, or closely related distributions, due to the computational issues that arise when one wishes to define interpolation nodes for general distributions. This paper examines the use of the recently introduced weighted Leja nodes for that purpose. Weighted Leja interpolation rules are presented, along with a dimension-adaptive sparse interpolation algorithm, to be employed in the case of high-dimensional input uncertainty. The performance and reliability of the suggested approach is verified by four numerical experiments, where the respective models feature extreme value and truncated normal parameter distributions. Furthermore, the suggested approach is compared with a well-established polynomial chaos method and found to be either comparable or superior in terms of approximation and statistics estimation accuracy.
Quantum amplitude estimation is a key sub-routine of a number of quantum algorithms with various applications. We propose an adaptive algorithm for interval estimation of amplitudes. The quantum part of the algorithm ...
详细信息
We study the arbitrary cost case of the unweighted Stochastic Score Classification (SSClass) problem. We show two constant approximation algorithms and both algorithms are 6-approximation non-adaptive algorithms with ...
详细信息
A novel second order family of explicit stabilized Runge–Kutta–Chebyshev methods for advection–diffusion–reaction equations is introduced. The new methods outperform existing schemes for relatively high Peclet num...
详细信息
Enriching Brownian motion with regenerations from a fixed regeneration distribution µ at a particular regeneration rate κ results in a Markov process that has a target distribution π as its invariant distributi...
详细信息
We propose a spectral method for the 1D-1V Vlasov-Poisson system where the discretization in velocity space is based on asymmetrically-weighted Hermite functions, dynamically adapted via a scaling α and shifting u of...
详细信息
暂无评论